Glove-based Tactile Streaming of Braille Characters and Digital Images for the Visually Impaired

为视障人士提供基于手套的盲文字符和数字图像触觉流传输

• 批准号: 10601900
• 负责人: Michel Albert Audette
• 金额: $ 25.72万
• 依托单位: HAPTX INC.
• 依托单位国家: 美国
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-09-30 至 2024-09-29
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10601900
• 关键词: 12 year old; 3-Dimensional; Address; Adult; Age; Algorithmic Software; Blindness; Braille Display; Calibration; Characteristics; Child; Collection; Communities; Computer software; Consultations; Coupling; Data; Development; Encapsulated; Endowment; Engineering; Eye diseases; Fingers; Foundations; Fracture; Genetic Transcription; Goals; Growth; Hand; Health; Image; Individual; Label; Length; Manuals; Maps; Mental Depression; Modality; Monitor; Multimedia; Network-based; Optometrist; Palpable; Paper; Patients; Perception; Phase; Population; Prevalence; Process; Quality of life; Readability; Reader; Reading; Research; Research Personnel; Resolution; Risk; Sculpture; Semantics; Stream; Students; Surface; System; Tablets; Tactile; Technology; Text; Thumb structure; Time; Validation; Vision; Visual Acuity; Visual impairment; age related; aging population; blind; braille; cohort; computerized; deep learning; deep neural network; demographics; density; design; digital; digital imaging; disability; fall risk; haptic feedback; haptic technology; haptics; indexing; lexical; multimodality; neural network; object recognition; open source; pressure; spatiotemporal; validation studies; virtual

This project will achieve a computerized system that supplies digital content, both lexical and graphical, suitable for children, students and working adults of the Visually Impaired (VId) community, to increase their autonomy and potentiate their Quality of Life and societal contribution. While there is notable academic research, such as a high-density pin display tablet, there is no commercial, broadly functional, full-page Braille reader. A tablet reader is potentially limiting in comparison with bimanual haptic gloves, limiting any content complementary to Braille text. In contrast, haptic gloves support both Braille lexical content and palpable digital images. The proposed system will encapsulate a full-length Braille reader, not limited to a page at a time but supplying a limitless stream of characters readable by four fingers in each hand as dictated by the content, while also supporting haptic actuation of embedded pictorial content. This glove-based approach will unlock, for an underserved VId clientele, a vast collection of multimedia content currently the exclusive domain of the seeing population, as Braille lexical content and palpable bas-relief of graphical content. Phase I will center on a haptic system for Braille display of full-length text, sensed by the fingers of both hands and based on a redesigned commercial bimanual glove-based haptic system. A software algorithm will stream and stagger the Braille content to each finger to replicate the bimanual reading process of a VId reader. The system will also feature a tactile bas-relief version of digital images embedded in the text. Phase II will develop Haptic VR, demonstrating the utility of overlaying Braille characters on a haptic curved surface and paving the way for vitally important haptic sensing of labeled 2D/3D graphics by VId users. A long-term goal is to apply this VR foundation to geolocated Haptic AR, enabling haptic overlay of Braille labels on everyday surfaces, from appliances to groceries, by suitable integration of haptic rendering with 3D range-sensing and DNN-based object recognition. Specific Aims – The Phase I Aims are as follows. Aim 1. High-resolution haptic gloves for Braille streaming- Development of bimanual high-resolution haptic gloves, featuring dense actuation at eight fingertips, by Haptx. The thumbs will also get haptic feedback. Aim 2. Haptic rendering of lexical and graphical content 2A. Lexical reader - A software transcription will support spatiotemporal streaming of lexical data that moves the haptic footprint of Braille characters serially from the first to the eighth finger. 2B. Graphical reader – The conversion of a digital image to a haptic bas-relief will exploit fast DNN-based image semantic segmentation, where each label will be mapped to an actuator pressure. 2C. Finger-based calibration- Both modalities will be supported by a virtual page defined on a tabletop, whose coordinates are found by a simple calibration based on three points identified by the index fingers.

该项目将实现一个提供数字内容（词汇和图形）的计算机化系统， 适合视障 (VId) 社区的儿童、学生和工作成年人，以提高他们的能力 自主权并提高他们的生活质量和社会贡献。 研究，例如高密度针显示平板电脑，没有商业化的、功能广泛的、整页的 与双手触觉手套相比，平板电脑阅读器可能存在局限性。 与盲文文本互补的内容相比之下，触觉手套支持盲文词汇内容和盲文内容。 可触知的数字图像。 所提议的系统将封装一个全长盲文阅读器，不限于一次一页，而是 根据内容的指示，提供每只手的四个手指都可以读取的无限字符流， 同时还支持嵌入图片内容的触觉驱动，这种基于手套的方法将解锁， 对于服务不足的 VId 客户来说，大量的多媒体内容目前是 视觉人群，如盲文词汇内容和图形内容的明显浅浮雕。 第一阶段将集中于触觉系统，用于盲文显示全文，由双方的手指感知 手并基于重新设计的基于商业双手手套的触觉系统。 将盲文内容流式传输并错开到每个手指，以复制 VId 的双手阅读过程 该系统还将在文本中嵌入数字图像的触觉浅浮雕版本。 第二阶段将开发 Haptic VR，展示在触觉上叠加盲文字符的实用性 曲面并为 VId 标记的 2D/3D 图形的极其重要的触觉传感铺平道路 用户的长期目标是将这个 VR 基础应用于地理定位触觉 AR，从而实现触觉叠加。 通过触觉渲染的适当集成，在从电器到杂货的日常表面上贴上盲文标签 具有 3D 距离感测和基于 DNN 的对象识别的具体目标 - 第一阶段的目标如下。 目标 1. 用于盲文流媒体的高分辨率触觉手套 - 开发双手高分辨率触觉 Haptx 的手套具有八个指尖的密集驱动功能，拇指也能获得触觉反馈。 目标 2. 词汇和图形内容的触觉渲染 2A. 词汇阅读器 - 软件转录将支持词汇数据的时空流 将盲文字符的触觉足迹从第一根手指连续移动到第八根手指。 2B. 图形阅读器——将数字图像转换为触觉浅浮雕将利用基于快速 DNN 的技术 图像语义分割，其中每个标签将映射到执行器压力。 2C. 基于手指的校准-两种模式都将由桌面上定义的虚拟页面支持， 其坐标是通过基于食指识别的三个点的简单校准来找到的。